Understanding The Risks In Credit Default Swaps

Transcription

1 STRUCTURED FINANCE Special Report AUTHOR: Jeffrey S. Tolk Vice President Senior Credit Officer (212) CONTACTS: Issac Efrat Managing Directior (212) Gus Harris Managing Director (212) Jeremy Gluck Managing Director (212) Investor Liaison Vernessa Poole All Asset Backed and Residential Mortgage Backed Securities (212) Sally Cornejo Collateralized Debt Obligations, Commercial Mortgage Backed, and Fully Supported Securities (212) WEBSITE: CONTENTS: Summary Opinion Credit Default Swaps The Typical Structure Isolating The Reference Portfolio's Credit Risk Moody's Definition Of Default And Loss ISDA Credit Events The Problem Of Soft Credit Events: Synthetic Vs. Cash Moral Hazard Marking-To-Market/Calculation Of Losses Reference Credits Other Than Corporate Obligations Good Faith Of The Sponsor Conclusion SUMMARY OPINION Credit derivatives offer unique opportunities and risks to investors. They allow investors to have exposure to a firm without actually buying a security or loan issued by that firm. Because the exposure is synthetic, the transaction can be tailored to meet investors needs with respect to currency, cash flow, and tenor, among other things. However, if the transaction is not structured carefully, it may pass along unintended risks to investors. Significantly, it may expose investors to higher frequency and severity of losses than if they held an equivalent cash position. Moody's has rated numerous structured transactions mostly synthetic collateralized debt obligations (CDOs) and credit-linked notes (CLNs) whose key feature is a cash-settled credit default swap. Under the swap, losses to investors are determined synthetically, based on credit events occurring in a reference portfolio. Investors risk, thus, is driven largely by the definition of credit events in the swap. The broader the definition, the broader the risk. The definitions published by the International Swaps and Derivatives Association (ISDA) are, in many respects, broader than the common understanding of default, and thus impose risk of loss from events that are not defaults. For example, Moody's and much of the market considers certain types of restructuring events to be defaults. However, the current ISDA definition of restructuring is broader than Moody's definition of default, and includes events that would not be captured by a Moody's rating. March 16, 2001

2 Likewise, the ISDA definitions for other credit events e.g., bankruptcy, obligation acceleration, and obligation default are broader than Moody's definition of default. For the Moody's rating of a reference portfolio to capture the risks to investors, the credit events should be narrowed such that they are consistent with defaults the events captured by a Moody's rating. Many of the risks in these transactions are driven by moral hazard the inherent conflict of interest that exists because the sponsoring financial institution (which is buying protection from investors) determines when a loss event has occurred as well as how much loss is imposed on investors. The sponsor's incentive, of course, is to construe credit events as expansively as possible and to calculate losses as generously as possible. Moody's considers these risks carefully when issuing its ratings. In addition to tightening the credit event and loss calculation provisions, these risks can be addressed by increasing transparency and providing mechanisms for objectively verifying loss determinations and calculations. Setting aside moral hazard, risks also arise based on the inherent difficulty in valuing a defaulted credit to determine the extent of loss to investors. Calculated losses may vary based on liquidity, market conditions, and the identity of the parties supplying bids. In analyzing a credit default swap, Moody's looks carefully at the methods and procedures for calculating loss given default, to ensure that all calculations are meaningful, realistic, and fair. The ISDA Credit Derivatives definitions, as currently drafted, do not effectively unbundle credit risk from other risks. If not structured carefully, a credit default swap using the ISDA definitions can pass along risks other than credit risk. For example, the swap may pass along the risk of loss following credit deterioration short of default. Such a risk is not necessarily captured by a Moody's rating of the reference portfolio, and, with some exceptions (e.g., when the loss event is a rating downgrade) is not readily capable of being measured. Table 1 Example of a Cash-Settled Credit Default Swap Seller Premium Credit Event Payment (100%-market value) x notional Reference Credit(s) Buyer The capital markets have an enormous capacity for absorbing credit risk, and this capacity has only been partially tapped by the credit derivatives market. In Moody s opinion, for capital markets investors to participate fully in the credit derivatives market, the risks inherent in credit default swaps must be more precisely defined, more transparently managed, and more readily quantifiable. This special report will describe the typical cash-settled credit default swap underlying a synthetic CDO or CLN, 1 compare Moody s definition of default with the credit event definitions currently used in the ISDA documentation, and discuss how different structural features and parameters of a credit default swap can affect risks to investors and impact Moody s analysis and ratings. CREDIT DEFAULT SWAPS Credit default swaps allow one party to buy protection from another party for losses that might be incurred as a result of default by a specified reference credit (or credits). The buyer of protection pays a premium for the protection, and the seller of protection agrees to make a payment to compensate the buyer for losses incurred upon the occurrence of any one of several specified credit events. 1 For purposes of this report, a CLN refers to a structure that references a single credit, and a synthetic CDO refers to a structure that references a portfolio of credits. 2

3 A swap can be structured to provide for either physical settlement or cash settlement following a credit event. In a physically settled swap, the buyer of protection delivers to the seller an obligation of the reference entity that has experienced a credit event. The seller pays par for that asset, thus reimbursing the buyer for any default-related loss that it would otherwise suffer. In a cash-settled swap, the buyer of protection is not required to deliver the defaulted credit, but values the credit for example, by marking it to market or by using a final workout value and is reimbursed for the loss (measured by the difference between par and the value following default). Because most of the synthetic CDOs and CLNs that Moody s rates are supported by cash-settled swaps, this special comment will focus on cashsettled credit default swaps. THE TYPICAL STRUCTURE Through synthetic CDOs or CLNs, financial institutions utilize credit default swaps to buy credit protection usually from the capital markets in the form of issued securities, but also directly from counterparties in the form of over-the-counter swap transactions. The structure allows financial institutions to remove credit exposure from their balance sheets while retaining ownership of the assets, and thus (1) manage risk more efficiently, and (2) obtain economic and/or regulatory capital relief. In the typical structure, the sponsoring financial institution (the entity seeking protection against credit losses) sets up a special purpose vehicle (SPV) to serve as counterparty to the credit default swap (making the SPV the provider of protection). The SPV is funded with the proceeds of notes issued to investors; it will use those proceeds to make credit event payments to the financial institution, and to return any remaining principal to investors at the deal s maturity. The proceeds of the securities are typically invested in highly rated securities in such a way that the ratings of the notes can be de-linked from the rating of the sponsoring institution. Under the swap, the SPV is the seller of protection, and the financial institution is the buyer. The swap references a credit exposure, or portfolio of credit exposures, for which protection is being provided. The arrangement is similar to an insurance policy, in which the financial institution is buying insurance against losses due to default in its portfolio. The credit exposures can be assets physically owned by the sponsor (e.g., loans, bonds, other securities), exposures to counterparties (e.g., by way of currency or interest rate swaps), or synthetic exposures (e.g., if the sponsor has sold protection on particular assets by way of credit default swaps). 2 Typically, in a synthetic CDO, the financial institution retains the firstloss piece, and the mezzanine tranches are securitized and sold to investors. There is often a supersenior piece that is either retained by the sponsor or passed off to a counterparty by way of a swap. Example of Cash-Settlement Under a Credit Default Swap Thirty days after a company defaults (e.g., it fails to make a bond coupon payment), its bonds are valued at 30 cents on the dollar. If the notional of a cashsettled credit default swap referencing that entity s bonds is $10 million, the protection seller would be required to make a payment of $7 million to the protection buyer. Notes Credit Default Swap Premium Investors SPV Sponsor Proceeds Table 2 Typical Structure Collateral Credit Event Payments Reference Credit(s) 2 However, there is no requirement that the sponsoring institution actually have exposure to the credits being referenced; it may merely wish to take a short position on the credit of the reference entities. 3

4 There are a number of key variations on the structure that can have a significant impact on the analysis of the transaction: The reference pool can be static remaining the same throughout the life of the transaction or it can be dynamic, permitting removal and substitution of the individual reference credits pursuant to portfolio guidelines. The swap can provide for ongoing cash settlement as defaults happen and losses are incurred or it can provide for cash settlement only at the maturity of the deal. The procedure and timing for determining severity of loss on a defaulted credit reference can vary from a bidding procedure that takes place shortly after a default, to reliance on a final work-out value established after the formal workout process has been completed. The swap can reference specific credits, or it can reference the general, unsecured debt of a reference entity. If the swap references the general, unsecured debt of an entity, credit events under the swap can be triggered by defaults only on bonds or loans, on a broader class of borrowed money, or on an even broader class of payment obligations. Perhaps most significantly, the definition of credit event can be tailored to meet the needs of the various parties to the transaction. While each of these variations is important, the most heavily negotiated component is most often the designation and characteristics of the credit events that will trigger a cash settlement under the swap. ISOLATING THE REFERENCE PORTFOLIO S CREDIT RISK Measurable Credit Risk Credit risk is generally viewed as the risk of loss following default. The risk of loss following other events other than default is not captured by credit risk. It is defined this way largely for practical reasons because default is an event that can be predicted based on historical data. Rating Downgrade as a Credit Event While Moody s does not have data concerning corporate level events (e.g., covenant breaches, accelerations) other than default, Moody s does have data on ratings transitions. With this data, it is possible to determine the likelihood of an entity transitioning from one rating category to another. For example, based on historical data, it is possible to determine the likelihood of a corporate obligor being downgraded from Baa3 to Ba1 over a certain period of time. Moreover, there is data concerning (1) spreads at the various rating categories, and (2) the widening of spreads following downgrade. Thus, it may be possible to assign a severity of loss, as well as a probability, to various downgrade events. Moody s has not been asked to rate a structure where the credit event is a rating downgrade, but it is theoretically possible to do so. A number of institutions have extensive data on defaults. Moody s, for example, has compiled data on corporate defaults spanning more than 80 years. This data permits Moody s to assign a probability of default to each rating category over a given time period. For example, based on historical data, Moody s can estimate that a firm rated Baa3 has a 6.1% chance of defaulting over a 10- year period. Moody s can also estimate the severity of loss given default for a given instrument using historical data. The market does not have comparable data for corporate events other than default. For example, Moody s does not have data concerning breaches of certain covenants, loan accelerations, or certain types of debt restructurings. 3 Thus, Moody s cannot assess the probability of these events occurring with the same accuracy that it can for defaults. Credit risk could be defined with respect to these events, but it would be very difficult to measure that risk without more data. 3 To date, requests made to other sources, such as banks and other financial institutions, have not yielded such data either. 4

5 Isolating the Risk The basic assumptions underlying a synthetic CDO or CLN are that (1) only the credit risk i.e., the risk of loss given default of the reference portfolio is passed through to investors, 4 and (2) the risk of loss on the reference portfolio can be measured by the rating of the portfolio. Thus, when analyzing a transaction, Moody s must confirm that these assumptions are correct: Credit risk must be truly isolated from other risks inherent in the reference portfolio. In its purest form, the swap should not transfer risks other than credit risk. Significantly, it should not pass through the risk of loss following credit deterioration short of default. The definition of credit event under the swap the event that requires valuing a reference credit and calculating a loss should be consistent with a well understood and measurable definition of default. For a Moody s rating of the portfolio to be used to assess the risk of the portfolio, the definition of credit event should be consistent with Moody s definition of default. In addition, the methods for calculating loss following default under the swap should be consistent with the market standard. If any of these assumptions are incorrect, a Moody s rating of the reference portfolio will not capture the risk to investors. For example, if the loss trigger events under the swap are broader than the events Moody s considers to be defaults, the actual expected loss posed to investors may be greater than the expected loss incorporated in the Moody s rating of the reference portfolio. Synthetic CDO and CLN investors may be subject to greater risks than if they actually owned the reference portfolio and held each reference asset to maturity. MOODY S DEFINITION OF DEFAULT AND LOSS In assigning ratings and compiling its historical default statistics, Moody s considers the following events to be defaults: Any missed or delayed disbursement of interest and/or principal; Bankruptcy or receivership; and Distressed exchange where (i) the borrower offers debtholders a new security or package of securities that amount to a diminished financial obligation (such as preferred or common stock, or debt with a lower coupon or par amount), or (ii) the exchange has the apparent purpose of helping the borrower avoid default. Severity of loss is defined as the difference between par and the recovery rate measured as a percentage of par following default. Moody s uses the market value of defaulted instruments, approximately one month after default, as an estimate of recovery rate. 5 These are the events that constitute defaults in Moody s historical studies, and these are the events that can be predicted by a Moody s rating. ISDA CREDIT EVENTS The 1999 ISDA Credit Derivatives Definitions 6 currently list six credit events that can be incorporated into credit swaps: Bankruptcy; Failure to pay; Restructuring, Repudiation/moratorium; Obligation default; and Obligation acceleration. 4 Transactions that are not de-linked from the sponsoring institution are also subject to credit risk associated with the sponsor. In assigning a rating to such transactions, Moody s will include sponsor credit risk in its analysis. 5 See Moody s Special Comment, Default and Recovery Rates of Corporate Bond Issuers: 2000, pp. 18, 24 by David Hamilton, Greg Gupton, and Alexandra Berthault (February 2001). 6 The ISDA Credit Derivatives Definitions have been revised once, and are currently in a state of flux. Thus, it is likely that they will be different two years from now. However, the 1999 definitions are currently the starting point for negotiations of most credit default swaps. Moreover, even though the definitions may change, it is fruitful to scrutinize them now to determine, as a general matter, what is and is not consistent with the Moody s definitions. For purposes of this special comment, familiarity with the ISDA definitions is assumed, and thus the definitions will be described only generally. 5

6 While these are the so-called standard credit events, their inclusion and scope are always heavily negotiated in the context of Moody s-rated synthetic CDOs and CLNs. The choice and characterization of these events is crucial, because they determine the probability of a loss occurring under the swap, as well as the extent of any such loss. Some of the ISDA credit events are consistent with Moody s definition of default, and some are not. Bankruptcy The definition of Bankruptcy in the ISDA Credit Derivatives Definitions was copied wholesale from the ISDA Master Agreement. Thus, while most of the definition is consistent with a default, there are some components that are not. The last clause of the definition, a catchall provision, is problematic because it makes a credit The ISDA Bankruptcy Definition The in furtherance of provision could be construed event any action by the reference entity in very broadly for example, the act of planning for, or furtherance of, or indicating its consent to, even considering, a bankruptcy filing (such as hiring approval of, or acquiescence in one of the listed bankruptcy advisors to discuss options) might be in bankruptcy events. This clause exposes investors furtherance of bankruptcy, but would not generally be to potentially greater risks, because it includes considered a bankruptcy event. 8 If publicly reported, events that are vague, difficult to identify, and do such exploratory steps could trigger a loss payment not clearly indicate default. 7 under the swap, even if the obligor does not ultimately Another potentially troublesome item in the ISDA enter bankruptcy. bankruptcy definition is insolvency. The ISDA definition does not specify what is intended by insolvency. However, there are different definitions for example, by reference to balance sheet or income statement tests and, depending on the definition used, the timing of an insolvency credit event could vary. Under a very broad definition, it is conceivable that an insolvency could occur without being followed by an actual bankruptcy or failure to pay. Thus, a broad interpretation could lead to a credit event being called under the swap when no default has actually occurred. Failure to Pay The ISDA failure to pay definition is consistent with Moody s definition of default. The key issue under this definition is materiality i.e., the missed payment should be in an amount that is material, such that it would be captured by a Moody s rating. To ensure that a credit event is not triggered by the failure to pay a trivial amount, a minimum amount referred to as the Payment Amount in the ISDA definitions should be specified under the swap. While there is a standard minimum amount, that amount may not be appropriate in all transactions, and it should be considered carefully for each swap. In some cases, the choice of a Payment Amount will depend on whether the swap is referencing (1) a specific obligation, (2) bonds or loans, (3) borrowed money, or (4) the more general payment obligations all of which are options under the current ISDA documentation. A Moody s rating will capture the risk of a failure to pay on the obligations rated by Moody s usually bonds and loans. However, it may not capture the risk of non-payment on all of an entity s payment obligations e.g., disputed trade obligations, certain fees, etc. An entity may choose not to make a payment on one of its payment obligations for reasons other than credit problems. To ensure that a Moody s rating will capture the risk of payment default, the category of obligations being referenced should be carefully considered. In some circumstances, a higher minimum payment amount may be appropriate. 7 It was not appropriate to use the entire bankruptcy definition from the ISDA Master for the credit derivatives definition, because the definition has different purposes in the different contexts. In the context of the ISDA Master, bankruptcy is an event that permits early termination of a swap - whether it be an interest rate swap, a currency swap, or any other kind of swap - because of credit problems with a counterparty. Naturally, to minimize their potential losses, swap participants want the ability to terminate a swap before a counterparty default actually occurs - e.g., when a counterparty takes any action in furtherance of a bankruptcy filing. By contrast, in the credit derivatives context, the purpose of the bankruptcy definition is to transfer default risk of a reference portfolio - not of the counterparty - from one counterparty to another, and not to give a counterparty the opportunity to terminate a swap early, before a default occurs. 8 See Lowenstein, Roger, When Genius Failed: The Rise and Fall of Long-Term Capital Management (Random House 2000), pp

7 Restructuring Moody s considers certain types of restructuring events known as distressed exchanges to be defaults, and captures those events in its ratings. Thus, Moody s does not believe that restructuring, as a concept, needs to be excluded from the credit derivatives definitions. In many respects, however, the current ISDA definition of restructuring is broader than Moody s definition of distressed exchange, and includes events that are not captured by a Moody s rating. 9 Thus, for a Moody s rating of the reference portfolio to capture the risk to investors, the definition of restructuring should be tightened to make it consistent with distressed exchange. Under the current ISDA restructuring definition, five events can qualify as a restructuring. Each event must meet the following requirements to qualify as a credit event: the restructuring (1) must not have been provided for in the original terms of the obligation, and (2) must be the result of a deterioration in the obligor s creditworthiness or financial condition. While these requirements are helpful in restricting the events that could constitute credit events, they are not sufficient to prevent overbroad applications of the definition. The first three events under the definition restructuring of an obligation that leads to (1) a reduction in interest payment amounts, (2) a reduction in principal repayment amounts, or (3) a postponement or deferral of interest or principal payments can constitute distressed exchange defaults under Moody s definition. Any one of these events, by itself, would arguably lead to a diminished financial obligation. However, if combined with other changes to the obligation, they may not. For example, an obligation that has been restructured to defer principal payments may not be considered a diminished financial obligation and thus not a distressed exchange if the lender has been compensated for the deferral. Thus, any restructuring definition should look at the totality of the circumstances e.g., whether the lenders/investors have been compensated for the reduction or deferral to determine whether the restructured obligation is truly a diminished financial obligation. Criteria for a Distressed Exchange Default Whether a distressed exchange default has occurred can be a subjective, fact-specific determination. In general, Moody s looks to three factors: (1) the extent to which the restructured obligation is a diminished financial obligation i.e., the degree of monetary impairment suffered by investors, (2) the extent to which the restructuring was involuntary for all investors, and (3) the extent to which the restructuring was done to avoid an imminent payment default. 10 A Real Life Example In August 2000, Conseco s bank debt was restructured. While the restructuring included a deferral of the loan s maturity by three months, it also included an increased coupon, a new corporate guarantee, and additional covenants in favor of the lenders. Thus, because lenders were compensated for the maturity extension, Moody s did not consider the restructured debt to be a diminished financial obligation, and thus not a distressed exchange default. (Indeed, the senior unsecured rating was downgraded only to B1). However, because of the maturity extension, the restructuring was considered a credit event under the ISDA restructuring definition and triggered loss payments under the credit default swaps written on Conseco. This is a perfect example of a loss event under the ISDA restructuring definition that Moody s did not consider to be a default. The fourth ISDA restructuring event a restructuring that leads to a change in an obligation s priority, causing it to be subordinated can be overbroad. The subordination of a debt obligation to equity or preferred stock would clearly be a default. (It would probably lead to a failure to pay as well thus, rendering this restructuring event unnecessary). However, a restructuring that merely lowers an obligation from a senior to a subordinated position in the capital structure (but not to equity) could also trigger a credit event under the current ISDA definition. In addition, the event could be triggered in a scenario where an entity takes on senior debt that effectively subordinates other debt in its capital structure. While these events may lead to a downgrade of the affected obligations, they would not necessarily render the obligations diminished financial obligations and thus not be a default in Moody s view The credit derivatives market is aware of the overbreadth of the current restructuring definition. The press has reported that swaps including this credit event often carry an additional premium to compensate protection sellers for the additional risk. 10 See Moody s Special Comment, Moody s Approach to Evaluating Distressed Exchanges, by David Hamilton and Sean Keenan (July 2000) 11 Other than removing it all together, bankers have addressed the potential overbreadth of this event by allowing it to be triggered only if the affected obligation is subordinated to equity status. There are undoubtedly other ways of addressing this problem as well. 7

8 Finally, the fifth event in the ISDA definition any change in the currency or composition of any payment of interest or principal is also potentially overbroad. Apparently, the provision was intended to apply to scenarios where a restructuring leads to debt being repaid in a currency that is non-convertible or highly volatile e.g., an emerging market currency that ultimately leads to the lender receiving an amount that is lower than the promised interest or principle. 12 Such a scenario might render the restructured obligation a diminished financial obligation. (Arguably, it would constitute a failure to pay as well). However, the definition would also include restructurings where the foreign exchange risk to the lender is hedged, or otherwise eliminated, such that the lender ultimately receives precisely what it was promised. Although such a scenario may not be a default, it would nonetheless trigger a credit event payment under the swap. Incorporating a notion of diminished financial obligation into the restructuring definition would address this problem as well. Any definition should also consider the extent to which the restructuring was involuntary to lenders/investors. Under the current ISDA definition, a credit event can be triggered if the lender voluntarily agreed to the restructuring (so long as it is the direct or indirect result of credit deterioration). However, Moody s might not consider such an event to be a default, even if it results in a diminished financial obligation. Bank loans are restructured quite frequently especially loans that are not syndicated and not all such events would be considered defaults. Voluntary Restructuring Not Necessarily a Default For example, at the request of a borrower whose credit has deteriorated somewhat, a bank might agree to less favorable terms (e.g., a lower coupon) on an existing loan. However, the primary motivation for the restructuring may be something other than credit: for example, the bank might want to preserve its business relationship with a borrower if the bank did not agree to the restructuring, the borrower might pay down the loan immediately and take its business elsewhere because it believes the borrower s credit will ultimately improve. This type of restructuring might not constitute a default, but would trigger a credit event under the current ISDA definitions. Credit risk is risk imposed on lenders by an obligor. It does not include, and a Moody s rating does not address, risks that lenders impose on themselves i.e., the possibility that lenders will voluntarily take losses that are not forced on them by obligors. Unfortunately, it is not easy to define involuntary, or to determine whether a restructuring was indeed involuntary. One possible means of capturing involuntariness or at least increasing the likelihood that a restructuring was indeed involuntary is to limit restructuring credit events to situations where there are a minimum number of lenders. Restructuring thus would only be available for reference obligations that are bonds, or syndicated loans with a minimum number of syndicate members. With a larger number of lenders (as opposed to a bilateral loan), it is less likely that an obligation would be restructured for non-credit reasons. 13 The third factor that Moody s considers in assessing distressed exchange defaults the extent to which the restructuring was done to avoid an imminent payment default also provides some indication of involuntariness. If a lender is faced with the choice of taking a potentially smaller loss now, by way of a restructuring, or a larger loss later, due to an actual payment default or bankruptcy, it is a fairly good indication that the restructuring was involuntary. Unfortunately, in many cases it will not be easy to determine whether there would have been a default but for the restructuring. Moreover, this standard goes to the obligor s motives in carrying out the restructuring a subjective factor that may not always be evident. 12 It is unclear what scenarios a change in composition of payments was intended to cover. It is difficult to imagine scenarios that would not already be included in a failure to pay or the first three events of restructuring. 13 In all cases, analysis of the Restructuring definition depends on whether the swap references a specific obligation or the general unsecured obligations of a reference entity. If it is the latter, there is potentially a much broader range of instruments that could be restructured - e.g., both bilateral and multilateral loans - triggering a credit event under the swap. The seller of protection - the investor - will not necessarily be familiar with all of the entity s outstanding obligations, and thus cannot assess the risk of a restructuring occurring with respect to each outstanding instrument. If the swap does not reference a specific obligation or class of obligations, the definition of the Restructuring credit event should be restricted considerably. For example, investors may wish to consider limiting the events within the Restructuring definition that can trigger a credit event. In addition, in all cases, to ensure that the restructuring is meaningful, a minimum restructured amount necessary to trigger a credit event - referred to as the Default Amount in the ISDA definitions - should be specified in the swap. 8

9 With Restructuring more than any other credit event, the need for a precise, objective definition is often at odds with the need for accuracy and predictability. Any objective definition risks being over-inclusive (e.g., Conseco), because the determination of whether a restructuring constitutes a default is often fact-based and subjective. However, it is possible to refine the current ISDA definition based on the factors discussed above such that it more accurately reflects a true default event. Repudiation/Moratorium Repudiation/moratorium was included in the ISDA definitions mainly to address actions by sovereign lenders, and thus, is not included in many synthetic CDO s, where the exposure is primarily to corporate credit. When applied to corporate credits, repudiation/moratorium is generally consistent with Moody s views of default although it is unclear how it would be different from failure to pay. However, there is concern with respect to the provision that includes as a credit event when a borrower challenges the validity of... one or more Obligations. This provision could be construed overbroadly to include situations where there is a legal dispute over a borrowing in which, for example, the borrower unsuccessfully challenges some terms of the borrowing that does not ultimately lead to a failure to pay interest or principal. Moody s would not necessarily consider such an event to be a default. In addition, if this event is to be included, the Default Amount, or minimum amount that can be subject to a repudiation in order to trigger a credit event, should be material, so that the repudiation of a trivial amount will not trigger a credit event. Obligation Default ISDA defines Obligation Default as a non-payment default i.e., a default other than a failure to pay that renders an obligation capable of being accelerated. Moody s has not been asked to rate a transaction that includes this credit event, and the market has moved away from including it. This is because the event is much broader than Moody s and most of the market s definition of default. Most bonds and loans contain representations, warranties, financial covenants, and non-financial covenants, the violation of which can give lenders the right to accelerate. While such violations can indicate credit deterioration (e.g., failure to maintain a minimum financial ratio, taking on additional debt, etc.), many such violations can be technical (e.g., failure to send a report). Of course, Moody s ratings do not capture the probability of a technical violation occurring. Moreover, even a covenant violation that represents serious credit deterioration would not be captured if the obligation is still current on interest and principal, and has not carried out a distressed exchange or become bankrupt. Moody s simply does not have data concerning such events that would allow it to assign a rating to them. Because inclusion of this event forces counterparties to mark-to-market an obligation before a payment default occurs, it will cause investors (i.e., sellers of credit protection) to take losses that they would not incur if they actually bought and held the obligation. For example, even though an obligation has suffered credit deterioration giving rise to a financial covenant violation, there is still a good chance that the obligation will pay both interest and principal in full. However, at the time of the violation, market bids will likely come in below par, because of concerns about the credit, or because of market sentiment, interest rate movements, or other systematic factors. Thus, while an investor that actually holds the obligation to maturity will get out whole, the investor selling protection will not. 9

10 Obligation Acceleration Obligation acceleration is similar to obligation default. However, to trigger a credit event, the non-payment default i.e., default other than a failure to pay must lead to a reference loan, bond, or other obligation actually being accelerated. Like obligation default, an acceleration, by itself, would not be captured by a Moody s rating. A failure to pay, bankruptcy, or distressed exchange following acceleration would be captured, but the acceleration itself would not. Acceleration is simply a lender s exercise of its contractual right, under certain circumstances, to declare a debt immediately due and payable. 14 As with obligation default, the events giving rise to a right to accelerate under obligation acceleration defaults other than a failure to pay are not considered by Moody s to be defaults and would not be captured by a Moody s rating. Consequently, a lender s decision to exercise its acceleration right following such events is not captured either. 15 There are three possible outcomes following an acceleration: (1) the borrower repays less than it owes (or becomes bankrupt), (2) the debt is renegotiated, or (3) the borrower repays everything that it owes. The first outcome is already captured by other credit events failure to pay and bankruptcy. The second outcome, depending on the circumstances, may be a distressed exchange restructuring. The third outcome the lender receives everything it is owed is not a default. Because the first and second outcomes are already captured by other credit events, and the third outcome is not a default, it is unclear what additional scenarios this credit event is intended to capture. 16 It has been suggested that the purpose of this credit event is timing i.e., because many accelerations are followed closely by either a payment default, bankruptcy, or restructuring, 17 including this event allows credit protection payments to be made earlier than they otherwise would. However, if the acceleration precipitates a true default, the default is likely to occur, at most, two or three months later, and it is difficult to justify why a counterparty cannot wait until it has suffered a true credit event to be compensated. More fundamentally, an acceleration where the lender receives everything it is owed clearly not a default would trigger a credit event under the ISDA definition. While historically rare, there have been instances of bond accelerations where investors have been paid par, thus leaving them with no loss. Moody s has not compiled its own data on such events, because they are not defaults. Moreover, while Moody s is unaware of any data with respect to accelerations of loans and private placements, anecdotal evidence suggests that acceleration followed by total recovery i.e., the lender gets all of its money back is more common. Inclusion of obligation acceleration as a credit event would not be as problematic if the market value of an obligation is always par when the lender will be fully paid off following acceleration. If that were the case, there would never be any loss following such credit events. However, it is very possible that the market value would come in at less than par even if the accelerated debt is fully repaid. 14 There may be circumstances where acceleration is automatic, but those would typically involve much more serious events - such as a bankruptcy or insolvency. 15 Obligation acceleration is the only ISDA credit event that can be fully at the discretion of the lender. (It is true that a firm s lenders could force it into bankruptcy; however, a mere bankruptcy filing is not a credit event - the filing must remain undismissed for 30 days, or a final judgment of bankruptcy must be entered). This is contrary to the principle that credit risk addresses loss events imposed by the borrower, and not a loss that the lender voluntarily assumes itself. 16 Forms of obligation default and obligation acceleration exist in the ISDA master agreement as early termination events, and it is possible that these provisions were merely copied from the ISDA master to the credit event definitions. However, as already noted, early termination events in the ISDA master are not intended to be defaults, but are intended to give a party the right to terminate the swap early if the counterparty has suffered serious credit deterioration but has not yet defaulted. 17 One theory behind acceleration, however, is that it is intended to permit the lender to get out whole. A lender can accelerate if, because of nonpayment defaults, it is no longer comfortable with the borrower s credit and wants to get all of its money back before there is an actual default. 10

11 For example, if only some of the borrower s outstanding debt is accelerated and the protection buyer is permitted to value the debt that remains outstanding, it is likely that the market value of the remaining debt will be valued at less than par. The acceleration is likely to have been triggered by some credit deterioration (though short of payment default), which would cause the remaining debt to trade at less than par. Even if the accelerated debt is valued, it is likely that a lender would agree to a grace period to allow the borrower to gather funds and make arrangements for paying down the debt, especially if the lender believed it could get par following acceleration. If bids on the accelerated debt come in during that grace period pursuant to the swap, it is possible that because of market or interest rate movements or, more likely, because of uncertainty about the credit itself the bids could come in below par, even if investors/lenders ultimately receive par following acceleration. 18 Another concern with obligation acceleration is that its inclusion as a credit event may create additional incentives. A protection buyer that accelerates a reference obligation will be reimbursed regardless of the outcome. Indeed, the buyer could get all of its money back on the loan and get additional compensation if bids on the non-accelerated debt come in at less than par. Because occurrence of the obligation acceleration credit event is often within the protection buyer s control, the additional incentive means that the event is more likely to occur in the presence of a swap than under normal circumstances. If the buyer has the right to accelerate, it is more likely to exercise that right if it can receive additional compensation for doing so. Thus, any historical data regarding the likelihood of an acceleration would probably understate the likelihood of its occurring when it is covered by a swap. 19 Obligation Acceleration and the Problem of Basis Risk Sponsors of synthetic CDO and CLN transactions can fall under two different categories: (1) those who are credit default swap end users, and (2) those who are not. The end users are buying protection on cash exposure to the reference Example of Increased Incentives A bank is considering whether to accelerate the loan of a borrower that has suffered some credit deterioration (and violated some covenants) but is not in imminent danger of defaulting on any payments. Consequently, the bank knows that it will get all of its money back if it accelerates. If the bank has bought protection on that borrower through a cash-settled credit default swap that includes obligation acceleration as a credit event, after it accelerates the loan, the bank could get bids on the remaining outstanding debt and get additional compensation. For example, if the remaining outstanding debt were bonds trading at 85 on the dollar, the bank will receive an additional 15% recovery. Without the swap, the recovery is 100%; with the swap, the recovery is 115%. The incentive to accelerate is obviously greater in the presence of the swap credits. In other words, they have actual exposure to the credits through loans or other business relationships with the obligors. Sponsors that are not end users are buying protection on synthetic exposure to the reference credits. They are exposed to the credits by way of credit default swaps i.e., they are selling protection on the credits to other counterparties, and if there is a credit event on those swaps they will be required to make a credit event payment to the other counterparties. End user sponsors recognize that obligation acceleration is not necessary, and, unless required to do so by regulators, have typically not asked for its inclusion their transactions. However, institutions that are hedging, or buying protection on, synthetic exposure have argued that inclusion of this event is necessary, because most of the swaps giving rise to their exposure include obligation acceleration. If the synthetic CDO or CLN does not include this credit event, there are potential loss events for which they are not hedged. Believing this additional risk to be significant, these institutions are often unwilling to take the incremental basis risk and have asked Moody s to rate the transactions with this event. 18 Obviously, if all of a borrower s debt is accelerated and everything is paid down quickly, there will be nothing to bid on anyway. 19 The additional incentives are more likely to be present in situations involving bilateral loans, as opposed to bonds or widely syndicated loans. 11

12 Moody s is reluctant to rate a credit event that is not a default and is only present because of a quirk in the ISDA definitions that has become standard. 20 The optimal solution is to remove obligation acceleration all together, or for it no longer to be standard. However, Moody s has been able to rate transactions including this event with the following modifications to the ISDA definition: Acceleration is only a credit event if, after the later of a minimum time period and the time to the next payment date on the obligor s obligations, the accelerated obligation has not been fully repaid. The rationale is that if the accelerated obligation is not fully repaid by the next payment date, it will likely never be fully repaid. Following acceleration, instead of cash settlement, the protection buyer delivers to the SPV an obligation of the reference entity (1) that has been accelerated, or (2) if the delivered obligation has not been accelerated, that matures earlier than the transaction matures. The SPV would only be permitted to sell the delivered obligation if it actually defaults; otherwise, it must hold it until it matures or is paid down. 21 This should remove the incremental market risk inherent in this event under a cash settled swap. Moody s has considered numerous alternatives to these solutions, and additional solutions may be acceptable. 22 However, the best solution would be to exclude this event all together. THE PROBLEM OF SOFT CREDIT EVENTS: SYNTHETIC VS. CASH Credit default swaps are intended to mimic the default performance of a reference obligation. Thus, for example, owning a CLN is often considered equivalent to having a cash position in the underlying reference obligation, except that the maturity, coupon, or other cash flow characteristics may be different. If an investor holds the CLN to its maturity, it should have the same risk of loss as if it held the reference obligation to its maturity. 23 Put another way, the CLN should only default if the reference obligation defaults. However, if so-called soft credit events Example: Synthetic Exposure Riskier Than Cash Position Two investors have exposure to the bonds of XYZ. The first investor actually owns the bonds. The second investor owns a CLN that references XYZ. Because of credit deterioration, XYZ violates some covenants on its outstanding bank debt, which leads its bank to accelerate the loan. The bonds, however, are not accelerated, but are trading at 85 on the dollar. The credit deterioration is not serious enough to lead to a default, and the bonds are ultimately paid off completely at maturity. Because the underlying swap includes obligation acceleration as a credit event (swaps typically allow a credit event to be triggered if a bond or loan has been accelerated), the CLN investor receives 85% of its par back and the CLN terminates. By contrast, the cash investor receives the coupon for the remaining term of the bonds and receives all of its principal back at the bond s maturity. events that are not truly defaults are included in the swap, that will not be the case. Selling protection through a cash-settled credit default swap e.g., owning a CLN (or a synthetic CDO) can actually be more risky than actually owning the reference obligation(s). This is because cash-settled credit default swaps essentially force investors to cash out of their position following a credit event. 24 If the swap includes credit events associated with credit deterioration short of default e.g., a broadly defined restructuring or obligation acceleration the CLN can default (the investor will receive less than the full par of the CLN) when the reference obligation has not. Thus, if a cash-settled credit default swap includes soft credit events, the investor may suffer losses that are not captured by a Moody s rating of the reference obligation(s), subjecting it to greater risk of loss than if it actually owned the reference obligation(s). 20 Fortunately, obligation default has not become a standard credit event. 21 An alternative to maturity matching, such as granting the SPV the right to put the delivered obligation at par, may have the equivalent effect of eliminating market risk. 22 When modeling the expected losses of a credit default swap containing obligation acceleration, it may also be appropriate to increase the probability of a loss occurring in the reference portfolio. However, because of the lack of historical data on this event (and the potential increased incentives for acceleration arising from the presence of swaps), the adjustment is likely to be fairly conservative. It would depend on a number of factors, including the nature of the reference obligations, the identity of the sponsor, the transparency of the transaction, and the motivations for doing the transaction. 23 Of course, if the investor chooses to sell the CLN prior to its maturity, it will be subject to market risk on that sale. 24 I.e., the seller must (1) mark the obligation to market, (2) make a cash payment equal to the difference between par and the market value, and (3) terminate the swap (or, in the case of a synthetic CDO, remove the affected reference obligation from the reference portfolio). 12